DIY GPU Machines

Building a GPU PC, Part 12: Installing the Graphics Card on the Motherboard

This time, we finally install the star of the show: the graphics card. 1. Remove the PCI Express slot covers The graphics card goes into the PCI Express Gen 5 x16 slot, so we remove the cover for that position. Also, this graphics card occupies 4 slots, so we remove

LLM Security

[LLM Security] Llama Guard: A First Step Toward AI Safety

Hello, this is the Qualiteg Research Department. Today we will explain Llama Guard, which was announced at the end of last year (December 2023). In recent years, AI technology has advanced remarkably and its range of applications continues to grow. At the same time, however, concerns about the safety of

AI-Business

[AI-Driven New Business Creation] Qualiteg's Original 7 Steps for Hypothesis Exploration and Customer Understanding

In new business development, clients are often unfamiliar with "hypothesis exploration." Before thinking up a business idea, you need customer problem hypothesis exploration to grasp your target users' real problems. This shapes service ideas the market truly wants and raises the odds of success.

DIY GPU Machines

Building a GPU PC, Part 11: Mounting the Motherboard and Wiring the Case

This time, we will mount the motherboard in the case and take care of the various cable connections. 1. Fit the standoff screws The case comes with what are called standoff screws (spacers). We will screw these into the case. The standoffs are there to hold the motherboard in place

AI Mathematics

Cross-Entropy Explained, Part 1: Two Formulas for One Loss Function

Good morning from the Qualiteg Research Department. Starting today, we would like to take a thorough look at cross-entropy. Are there two formulas for the cross-entropy function? In this series, we take up the cross-entropy function, which is commonly used as the loss function for classification problems in machine learning.

Daily Dev Tips

Buffering Incrementally Generated Tokens

Hello from the Qualiteg Product Development Department. Today we would like to introduce a way to deal with the fragmented-token problem that often comes up during inference. In the incremental generation used for streaming chat, text is generated one token at a time. A token, however, is neither a "

AI-Business

[AI-Driven New Business Creation] Qualiteg's Original Guide to Framing a New Business Direction and Building Consensus with the Business Owner

How to build consensus on goals with a new business's owner. Step 1 clarifies the vision and objectives and sets numerical targets. Step 2 creates a roadmap and KPIs to reach them. Step 3 shares progress and adjusts through regular meetings, building an effective structure to drive the business.

DIY GPU Machines

Building a GPU PC, Part 10: Installing the CPU, Memory, and SSD on the Motherboard

In this installment, we set up the motherboard and its main components. 1. Unboxing the Motherboard The motherboard we purchased this time is the ASRock Z690 Steel Legend WiFi 6E/D5. Let's open it up right away. When you open the box, the cables and manuals are stored

Non-LLM Generative AI

Trying Out MagicAnimate

Magic Animate is an open-source project that turns still images into videos. In this article, we try it out using the Windows version. The test environment is a Core i7 + 64 GB RAM + RTX 3090. Note that installation takes about 30 minutes and consumes roughly 32 GB of storage. Creating

ChatStream Guide

[ChatStream] ChatPrompt for lightblue/karasu-7B-chat-plus

Hello, this is the Product Development Team at Qualiteg. In this article, we introduce the ChatPrompt for lightblue/karasu-7B-chat-plus. Customers using an older version of ChatStream can use this model by importing this ChatPrompt. (It is bundled in the latest release.) from chatstream import AbstractChatPrompt from chatstream.chat_prompt.role_

AI Mathematics

AI Math Basics: Logarithms and Exponentials, Part 3: Differentiating Logarithmic Functions (Continued)

Good morning. This is the Research Team at Qualiteg. Today we present the second half of our discussion on differentiating logarithmic functions. By the end of this installment, you should have a solid grasp of how to differentiate logarithmic functions. 2. Logarithm Identities First, let's review the basic

AI-Business

[AI-Driven New Business Creation] Qualiteg's Approach to Quantitative Goal Definition for Proving Your Own Results

Many clients set no numerical targets for a new business and struggle later. Quantitative goal definition sets objectives in concrete numbers so progress can be measured clearly, boosting motivation, resource allocation, and results. This column explains why numerical targets matter early on.

ChatStream Guide

[ChatStream] ChatPrompt for Llama 2

Hello, this is the Product Development Department at Qualiteg. In this article, we introduce the Llama 2-compatible ChatPrompt that is now bundled with ChatStream. The current ChatPrompt is shown below. If you are using an older version of ChatStream, you can add Llama 2 support with the following code. (Naturally,

DIY GPU Machines

Building a GPU PC, Part 9: Parts for an AI Research GPU Machine and Starting the Build

Yesterday I went to Akihabara and bought a total of about 415,000 yen worth of parts for a GPU machine, so here is a report. GPU Machine Parts * Motherboard: ASRock Z690 Steel Legend approx. 40,000 yen * Memory: Crucial desktop DDR5 32GB x 2 = approx. 30,000 yen * CPU:

ChatStream Guide

[ChatStream] ChatPrompt for rinna/nekomata-14b-instruction

In this article we introduce a ChatPrompt for rinna/nekomata-14b-instruction, which was released on December 21, 2023. The nekomata series is based on Qwen and has a vocabulary size of 152,000, considerably larger than earlier models, so we have high expectations for its Japanese language capabilities. This time we

ChatStream Guide

[ChatStream] Adding a Progress Bar to Slow Model Loading

Hello, this is the Product Development Division at Qualiteg. Loading an LLM from Hugging Face can take a very long time. In this article, we introduce a handy tool for exactly that situation. When you download a Hugging Face LLM, you can see the progress, but once the model has